I'm planning to release source code/snippets of my research. Please understand that the work consists of multiple large entangled systems with ten of thousands lines of code, so this will take time and I may not be able to provide support. The "research-code" is neither polished nor properly commented, but I decided to release it now for research and educational purposes. No commercial use allowed.
Here's the network training code that takes as input processed MFCC coefficients and mouth fiducial points represented as PCA coefficients.